So I already have a replicaset up and running and I want to add a deployment that'll manage that replicaset. I expect to go through this process repeatedly, so a programmatic solution is desired.
Right now my best solution is to run
JSON=kubectl get replicasets myreplicaset -o json
IMAGE=...
REPLICAS=...
LABELS=...
kubectl run --image=$IMAGE --replicas=$REPLICAS --labels=$LABELSwhere ... is json-chopping code.
I'm worried about the reliability of this code. Perhaps I am forgetting important parameters in the replicaset. I definitely want to avoid accidentally altering the replicaset (or having the deployment scale it down and launch a new one). Is there a simpler way to achieve the same effect?
You can try it with this command:
sed 's/ReplicaSet/Deployment/g' /tmp/replicaset.yaml | sed -n '/status:/q;p'> Deployment.yaml
it worked for me. but i would always recommend to apply deployments in the first place and not replicasets.